TRIPOD-Code is an extension to the existing TRIPOD (Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis) reporting guideline, focused on improving the reporting and transparency of code associated with prediction model studies. The scope of TRIPOD-Code includes all studies that develop, validate, or update multivariable prediction models for diagnostic or prognostic purposes, using any statistical or machine learning method. The primary aim is to provide structured guidance to ensure that the code underlying these studies is shared in a way that supports reproducibility, transparency, and reuse.While the TRIPOD and TRIPOD-AI guidelines encourage the sharing of analytical code, they do not provide detailed guidance on what constitutes sufficient or complete code sharing. When code is made available, it is often incomplete, undocumented, or difficult to reuse. This limits the ability of other researchers to verify results, assess methodological quality, or implement models in clinical or research settings.TRIPOD-Code addresses this gap by providing a checklist of minimum reporting items specifically focused on code availability, completeness, and usability. These items are expected to include documentation of software dependencies, licensing terms, file structure, testing procedures, and versioning or archival strategies.By developing TRIPOD-Code, we aim to help improve the quality and accessibility of code in prediction model studies.
Registered: 30 October 2025
Last updated:
Contact: Tom Pollard Email: [email protected]