Blog from December, 2017

Earlier this week, the curators released the metadata for SDTM v1.6 Final (https://www.cdisc.org/standards/foundational/sdtm) and SENDIG-DART v1.1 Final (https://www.cdisc.org/standards/foundational/send) to CDISC SHARE Exports (https://www.cdisc.org/members-only/share-exports). DART, an acronym for Developmental and Reproductive Toxicology, refers to study data typically found in embryo-fetal developmental toxicity studies. CDISC SHARE Exports is a part of the CDISC website where CDISC members can freely access static metadata output from SHARE. Please contact our Membership team for assistance with access (https://www.cdisc.org/contact).

Released together with the aforementioned metadata is Diff report. Diff, a shorthand for differences, is a computing term whose main purpose is to detect and display the alterations between two files. Changes made to SDTM v1.6 from its predecessor such as new variables and updated properties are detailed in the report. Another Diff report is comparing v1.1 to v1.0 of SENDIG-DART. SENDIG-DART v1.0 was published in August, 2016 with a Provisional status. This second Diff is a useful tool to aid patching or updating to the Final version.  

Curation for SENDIG-DART's metadata was particularly challenging; the SENDIG-DART is intended to be used in close concert with the SENDIG v3.1. The SENDIG-DART supplements the SENDIG with seven (7) new domains, spanning special purpose, findings, and trial design classes. Majority of the changes, however, lies in Section 7 - Changes to Existing Domains (https://www.cdisc.org/system/files/members/standard/foundational/send/SENDIG%20DART%20v1.1.pdf#page=38) and Appendix C - Repro Phase Day Timing Variables (https://www.cdisc.org/system/files/members/standard/foundational/send/SENDIG%20DART%20v1.1.pdf#page=45). These sections document modifications to sixteen (16) other domains, of varying degrees, for DART studies. The modifications are written primarily for human consumption, which poses the challenge. Extra time was necessary to carefully curate the full domain specification for those affected domains. Curation mostly entailed interpreting textual information in multiple subsections (e.g., “After RFENDTC, before RFXENDTC”, “BWBLFL variable is Permissible for DART studies”, etc.), then stitching them back to domain specification tables in SENDIG v3.1.

To ensure accurate interpretation and complete coverage, the curators arranged a special quality review cycle with the SEND Leadership Team. During which, not only a few corrections were advised, but a last-minute update to the SENDIG-DART document was also warranted. The entire curation process took about three and half months.

The metadata posted onto CDISC SHARE Exports would be impossible without the eyes and expertise by all parties involved. The end result is worthwhile as fully vetted metadata (Excel and Define-XML v2.0 for SENDIG-DART) are accessible to all CDISC members. They are a significant time saver to otherwise a tedious process to be undertaken by each sponsor company.

On a different note, it is opportune to summarize other metadata-related work completed in 2017. The SHARE team conducted a public input for both SHARE 2.0 Model (Q3) and SHARE 2.0 API (Q4). They are important work to lay a strong foundation for delivery CDISC metadata in a machine-readable format. They are the backbone for structured metadata, tools, and automation so that CDISC will be able to make new metadata available much more rapidly. We are starting to reap benefits from several homegrown tools: 1) CRF Maker that uses metadata to automate CRF annotations for use in Therapeutic Area User Guides; 2) QRS Maker that generates questionnaire dataset with both question and result metadata; 3) Spec Grabber that swipes properly structured dataset specification tables in CDISC WIKI and produces machine-readable metadata. This is an area we will certainly do a lot more in 2018.

For coming attractions, there are a few development work this community may anticipate during Q1 and Q2 next year. Sample metadata will be loaded into our API specification development platform, hosted on Swaggerhub (https://app.swaggerhub.com/apis/CDISC1/share-2.0). This will greatly enhance the API testing experience, especially for prospective users to experiment with RESTful API. A SHARE 2.0 soft launch, both Model and API, is projected for late Q2 with a good set of popular versioned standard included. If time permits, HL7 FHIR mappings to CDASH and SDTM, limited to concepts covered in CDISC’s Diabetes Therapeutic User Guide, could be in scope. More details about the soft launch will be shared once they become available.

Lastly, thank you for your partnership in 2017. We truly believe CDISC members drive global standards.

May you have a restful holiday and a rewarding new year to come.