SDTM Trial Summary (TS)

Description

A trial design domain that contains one record for each trial summary characteristic. This domain is not subject oriented.

Specification

Assumptions

1. The intent of this dataset is to provide a summary of study/trial information. This is not subject-level data.
2. Recipients may specify their requirements for which study/trial summary parameters should be included under which circumstances.
3. The order of parameters in the examples of TS datasets should not be taken as a requirement. There are no requirements or expectations about the order of parameters within the TS dataset.
4. The method for treating text >200 characters in TS is similar to that used for the Comments (CO) special-purpose domain. If TSVAL is >200 characters, then it should be split into multiple variables, TSVAL-TSVALn.
5. A list of values for TSPARM and TSPARMCD can be found in CDISC Controlled Terminology.
6. Controlled terminology for TSPARM is extensible. The meaning of any added parameters should be explained in the metadata for the TS dataset.
7. For a particular study/trial summary parameter, responses (values in TSVAL) may be numeric, datetimes or amounts of time represented in ISO8601 format, or text. For some parameters, textual responses may be taken from controlled terminology; for others, responses may be free text.
8. For some study/trial summary parameters, CDISC Controlled Terminology includes codelists for use with TSVAL. Recipients may also specify controlled terminology for TSVAL. These specifications may be for study/trial summary parameters for which there is no CDISC Controlled Terminology or they may replace CDISC Controlled Terminology for a study/trial summary parameter.

9. There is a code value for TSVALCD only when there is controlled terminology for TSVAL. For example, when TSPARMCD = "PLANSUB" (Planned Number of Subjects) or TSPARMCD = "TITLE" (Trial Title), then TSVALCD will be null.

10. TSVALNF contains a “null flavor,” a value that provides additional coded information when TSVAL is null. For example, for TSPARM = "AGEMAX" (Planned Maximum Age of Subjects), there is no value if a study does not specify a maximum age. In this case, the appropriate null flavor is "PINF", which stands for "positive infinity." In a clinical pharmacology study conducted in healthy volunteers for a drug where indications are not yet established, the appropriate null flavor for TSPARM = "INDIC" (Trial Disease/Condition Indication) would be "NA" (i.e., not applicable). TSVALNF can also be used in a case where the value of a particular parameter is unknown.
11. Some codelists used for TSVAL include terms which are also null flavors. For example, the Pharmaceutical Dosage Form codelist includes the values "UNKNOWN" and "NOT APPLICABLE". In such cases, TSVAL should have the term from the codelist and TSVALNF should be null.

12. For some studies/trials, there will be multiple records in the TS dataset for a single parameter. For example, a study/trial that addresses both safety and usability testing could have 2 records with TSPARMCD = "TTYPE" (Trial Type), one with the TSVAL = "SAFETY" and the other with TSVAL = "USABILITY TESTING". TSSEQ has a different value for each record for the same parameter.

    Note that this is different from datasets that contain subject data, where the --SEQ variable has a different value for each record for the same subject.

13. TS does not contain subject-level data, so there is no restriction analogous to the requirement in subject-level datasets that the blocks bound by TSGRPID are within a subject. TSGRPID can be used to tie together any block of records in the dataset. TSGRPID is most likely to be used when the TS dataset includes multiple records for the same parameter.

    For example, if a trial compared administration of a total daily dose given once a day to that dose split over 2 administrations, the TS dataset might include the following records.  There are 2 records each for TSPARMCD = "Dose" and TSPARMCD = "DOSFREQ".  Records with the same TSGRPID are associated with each other. In this example, dose units are the same for both administration schedules, so only 1 record for DOSU is needed.

    $titleHtml

    TSSEQ	TSGRPID	TSPARMCD	TSPARM	TSVAL
    1	A	DOSE	Dose per Administration	50
    1	A	DOSFREQ	Dosing Frequency	BID
    2	B	DOSE	Dose per Administration	100
    2	B	DOSFREQ	Dosing Frequency	Q24H
    1		DOSU	Dose Units	mg

    $warningHtml

14. Protocols vary in how they describe objectives. If the protocol does not provide information about which objectives meet the definition of TSPARM = "OBJPRIM" (Trial Primary Objective; i.e., the principal purpose of the study/trial), then the objectives should be provided as values of TSPARM = "OBJPRIM". Consult the controlled terminology for trial/study summary parameters for appropriate parameter values for representing other objective designations (e.g., secondary, exploratory).
15. As per the definitions, the primary outcome measure is associated with the primary objective, the secondary outcome measure is associated with the secondary objective, and the exploratory outcome measure is associated with the exploratory objective. It is possible for the same outcome measure to be associated with more than 1 objective. For example, 2 objectives could use the same outcome measure at different time points, or using different analysis methods.
16. If a primary objective is assessed by means of multiple outcome measures, then all of these outcome measures should be provided as values of TSPARM = "OUTMSPR" (Primary Outcome Measure). Similarly, all outcome measures used to assess secondary objectives should be provided as values of TSPARM = "OUTMSSEC" (Secondary Outcome Measure), and all outcome measures used to assess exploratory objectives should be provided as values of TSPARM = "OUTMSEXP" (Exploratory Outcome Measure). Additional key measures of a study that are not designated as primary, secondary, or exploratory should be provided as values of TSPARM = "OUTMSADD" (Additional Outcome Measure).
17. Trial indication: Values for TSVAL when TSPARMCD = "INDIC" would indicate the condition, disease, or disorder the study/trial is intended to investigate or address. A vaccine study of healthy subjects, with the intended purpose of preventing influenza infection, would have TSVAL = "Influenza". A clinical pharmacology study of healthy volunteers, with the purpose of collecting pharmacokinetic data, would have no trial indication; TSVAL would be null and TSVALNF = "NA" if TS contains a row where TSPARMCD = "INDIC".
18. Values for TSVAL when TSPARMCD = "REGID" (Registry Identifier) will be identifiers assigned by the registry (e.g., ClinicalTrials.gov, EudraCT).