The short version
- The Trial Summary (TS) dataset holds one row per study-level parameter: design, population, treatments, dates and the standards used.
- FDA uses the study start date in TS to decide which data standards requirements apply, and its technical checks expect a TS file to be present.
- A fifteen-minute review of TS before every transfer to the submission team avoids one of the most avoidable rejections.
What TS is
Most SDTM datasets describe subjects. TS describes the study. Each row is a parameter, identified by a short code in TSPARMCD with its value in TSVAL: the planned number of subjects, the trial phase, the control type, the indication, the study start date and many more [1]. Many values come from controlled terminology, so a reviewer, or a program, can read the design of the study without opening the protocol.
Because TS is small, it is often built last, by whoever has time, from a protocol synopsis that has since been amended. That is exactly backwards for a dataset that is checked first.
Why it matters more than its size suggests
FDA's requirement to submit study data in the standards listed in its Data Standards Catalog applies according to when a study started [2]. The study start date in TS is the evidence. For this reason FDA asks for a TS dataset with the study start date even for studies that started before the requirement, and the Study Data Technical Conformance Guide describes a simplified TS file for that case [3].
FDA's technical rejection criteria for study data, now part of the Technical Conformance Guide, turn this into an automated gate. The eCTD validation specifications list the checks, their effective dates and their severity [4]. Among the high-severity checks are the presence of a TS file with the study start date for each study in the relevant sections, and the presence of the demographics dataset and define.xml for tabulation data, or ADSL and define.xml for analysis data. A submission that fails a high-severity check can be rejected before any reviewer looks at the science.
Note the definition FDA uses: for clinical studies the study start date is the earliest date of informed consent for any subject enrolled in the study. It is not the date the protocol was approved, the first site was activated or the first subject was dosed. Using the wrong one is a quiet error that only surfaces when someone compares TS with the data.
A fifteen-minute TS review
Run this before each transfer to the submission team, not only before the final one.
| Check | How | Typical finding |
|---|---|---|
| Study start date present | Look for TSPARMCD = SSTDTC | Missing in a legacy conversion |
| Start date correct | Compare with the earliest informed consent date in DS or DM | First dosing date used instead |
| Design matches the current protocol | Planned subjects, arms, control type, blinding | Values copied from protocol version 1 |
| Terminology valid | Validate coded values against the declared terminology version | Free text where a code is expected |
| Standards declared | Parameters for the SDTM and terminology versions used | Versions differ from define.xml |
| One study identifier | STUDYID matches every other dataset and the eCTD study tagging | A different identifier in TS |
A synthetic example
The fictional ZEPHYR-101 team converts a legacy extension study for an integrated summary. The converted TS has no study start date, because the legacy database never stored one. The programmer derives it from the earliest informed consent date in the converted disposition data, documents the derivation in the reviewer's guide and adds the TS parameters for the versions of SDTM and controlled terminology actually used. The whole fix takes an hour in October. Found by an automated gate in a submission window, it would have cost the submission date.
Do this today
- Open the TS dataset of your next study to be transferred and confirm that
SSTDTCexists and equals the earliest informed consent date. - Compare the TS design parameters with the current protocol amendment, not the original synopsis.
- Check that the standards versions declared in TS match those declared in define.xml.
Check yourself
One question
Show the answer
Answer 2. The requirement to submit standardized study data depends on when a study started. The study start date in TS is how that is determined, which is why FDA asks for a TS file carrying it even for studies that are not otherwise required to use the standards.
Apply it
Turn it into a plan
The brief says what to check. The Navigator and the Handbook turn it into dated, owned actions for your milestone, in your browser, with no study data.
Sources
References
- CDISC. Study Data Tabulation Model Implementation Guide (SDTMIG) version 3.4. www.cdisc.org ↗
- U.S. Food and Drug Administration. Providing Regulatory Submissions in Electronic Format: Standardized Study Data. Guidance for industry. www.fda.gov ↗
- U.S. Food and Drug Administration. Study Data Technical Conformance Guide: Technical Specifications Document. www.fda.gov ↗
- U.S. Food and Drug Administration. Electronic Common Technical Document (eCTD): submission standards, including the Specifications for eCTD Validation Criteria. www.fda.gov ↗
Standards and guidance change. Confirm the version in force for your study, sponsor and region before relying on any reference. Examples are synthetic and illustrative. This brief is generalized professional guidance, not regulatory advice.
Next brief · Tuesday 10:00 ET · Lock Room