In clinical trials, the accuracy and integrity of data are paramount. While the goal is to handle data systematically and programmatically, there are occasions when hardcoding becomes necessary. Below are considerations for when to hardcode and the importance of documenting these decisions. 

When to Consider Hardcoding 

Hardcoding in clinical trial data is generally discouraged and should be a last resort because it introduces a risk of human error and may reduce the reproducibility of analyses. Typically, hardcoding is appropriate in scenarios when an error to subject data is identified after the database has been locked and when time constraints prevent reopening the database.  

Best Practices for Performing Hardcoding 

When hardcoding is necessary, transparency is essential to maintain the integrity of your data: 

1. Isolate the Hardcoding: Create a separate section in programs for hardcoding that is clearly commented for why hardcoding is being done, referencing relevant documentation. 
2. Minimize Hardcoding: Limit the scope of hardcoding to the smallest necessary changes and avoid hardcoding values that could be derived programmatically. 
3. Peer Review: Review the hardcoding for accuracy and necessity through double independent programming. 

Documenting Hardcoding 

Proper documentation is critical when hardcoding data in clinical trials. Every organization should have a clear and standardized procedure for documenting hardcoded changes to ensuring transparency during audits or regulatory reviews. This documentation should be signed and filed, at a minimum, within the study Trial Master File. 

What to Include when Documenting the Hardcoding: 

1. Description of the Issue:
   • Explain the original data issue, why it occurred, and why hardcoding was necessary.
   • Reference any communications or decisions that led to the hardcoding.
2. Details of the Hardcoding:
   • Provide a detailed description of what was changed.
   • Specify which datasets or variables were affected.
3. Impact Assessment:
   • Discuss the potential impact of the hardcoding on the analysis and results.
   • Confirm that the hardcoding does not affect the overall study integrity.
4. Review and Approval:
   • Include signatures or approvals from the necessary personnel, such as the study sponsor, study lead, data manager, and biostatistician.

Conclusion 

While hardcoding in clinical trial data should be a last resort, understanding when and how to do it properly is crucial for maintaining data integrity. Questions about how to handle a data issue you’ve encountered? Contact us to speak with one of our Data Standards experts. 

Eric Howard, Principal Statistical Programmer, has more than a decade of programming expertise, specializing in the creation and validation of CDISC standardized datasets (SDTM and ADaM), conducting statistical analyses, and ensuring the accuracy of data displays. His experience spans various therapeutic areas, covering studies from pre-clinical stages to phase 1 through 3 trials. Mr. Howard effectively communicates with sponsors to align expectations and facilitate informed decision-making and has successfully led teams in developing and organizing high-quality statistical outputs for NDA, DSMB, ISS, and ISE submissions. He holds a bachelor’s degree in Statistics and a master’s degree in Biostatistics from Grand Valley State University.