Dataset-JSON as Alternative Transport Format for Regulatory Submissions

The scope of this project covers twin objectives:

1. Establish Dataset-JSON as an acceptable drop-in replacement for XPT.
2. Position Dataset-JSON as the natural successor for XPT with advanced capabilities.

NOTE that deprication of XPT and adoption of advanced Dataset-JSON is a long-term goal and outside the scope of this project.

To this end, the scope of the project can be split into four:

Industry Adoption

• High-level enagagement with industryHigh-level engagement with industry
• Selling the sizzle! The ‘why’ not the ‘how’
• Making the business case for the sceptics, not only the early-adopters

Regulatory Acceptance

• Building confidence one small step at a time
• Dataset-JSON as drop-in replacement for XPT
• Testing, proof-of-concepts, pilots

Standards Development

• Long-term vision and philosphy
• Moving beyond limitations of XPT
• Maintaining backward compatability

Technical Implementation

• Education based on ‘just use the tools’
• Hackathons, hands-on experience and documentation
• Open-source, language-agnostic

The SAS Transport v5 (XPT) format dates from 1989 and was first available as part of SAS version 5. Since that time, there have been many changes to the industry with respect to the process for submissions and the approaches to data curation and manipulation – but none to the format itself.

The XPT specification is controlled and publicly documented by SAS Institute, who consider it a non-proprietary format.

Dataset-JSON is an alternative, modern dataset format that addresses the limitations of XPT and is published as part of the CDISC ODM v2 standard under an open-source MIT licence.

The challenge in moving from XPT to Dataset-JSON is twofold:

1. Ensure regulatory authorities have sufficient confidence in the Dataset-JSON to be willing to accept it as an alternative to XPT.
2. For industry to be able to transition to Dataset-JSON with no impact on the submission process.

The XPT file format is outdated and imposes limitations on submission data due to the lack of modern features – for example, XPT files are limited to US ASCII Character encoding, have restricted variable names/sizes, inefficient use of storage space, use a non-extensible tabular structure and lack any integrated metadata.

The impact of overcoming these limitations will be seen initially by regulators; however, by far the bigger impact will be seen in the medium to longer term, where, by lifting the restrictions of XPT, and moving to Dataset-JSON, new standards will be able to take advantage of these enhanced capabilities, and drive efficiencies, consistency and re-usability across the clinical research data life cycle.

Sam Hume (CDISC)

shume@cdisc.org 

Stuart Malcom (Veramed)

stuart.malcolm@veramed.co.uk

Jesse Anderson (FDA)

jesse.anderson@fda.hhs.gov

lauren@phuse.global 

PHUSE Team formation 

Engaged with collaboration partners (CDSIC, FDA etc.)

Q2 2023

Initial technical/ regulatory deliverables

CSS workshop – Workstream roundtable 

All project deliverables published 

CSS workshops – Project findings and recommendations

New project approved. Calling for volunteers.