Comparing Analysis Method Implementations in Software (CAMIS)
Project Description |
The CAMIS (Comparing Analysis Method Implementations in Software) project builds on the earlier Clinical Statistical Reporting in a Multilingual World initiative, addressing differences observed in statistical results across programming languages. Even within validated environments, variations in underlying computational approaches can produce results that differ but remain statistically valid. These discrepancies create uncertainty for sponsor companies when submitting to regulatory agencies, particularly regarding how such differences will be interpreted. CAMIS aims to define this challenge and provide a structured framework for assessing cross-language differences in statistical analyses. The project focuses on key analyses commonly used in submissions (e.g. summaries, models, bioequivalence testing), documents where equivalence can be achieved and evaluate known differences through practical use cases. It also provides sample code via a public repository and promotes statistically sound implementation choices over software-specific defaults, supporting greater confidence in multi-language regulatory submissions. The CAMIS repository to document known differences is now live and open for community contributions. |
Project Leads | |
|---|---|
Lyn Taylor, Parexel | |
Christina Fillmore, GSK | |
Yannick Vandendijck, Johnson & Johnson | |
Nicola Newton, PHUSE Project Coordinator | nicky@phuse.global |
Click above for the latest project status. |