Understanding SDTM Trial Summary Domain: ACTSUB and Screen Failures

In SDTM, the Trial Summary (TS) domain captures study-level information about a clinical trial. It does not contain subject-level data, but instead stores key summary values such as trial characteristics and subject counts.

TS is commonly included in SDTM-based submissions to provide high-level trial information.

One parameter that often creates confusion is TSPARMCD = ACTSUB. It is sometimes interpreted as the number of treated subjects, but that interpretation is not always correct.

What Is the TS Domain?

The TS domain contains trial-level metadata. Each record represents a parameter:

TSPARMCD = parameter code
TSPARM = parameter description
TSVAL = parameter value

TS values should be consistent with the study protocol, metadata, and define.xml.

What Does ACTSUB Mean?

ACTSUB is the TS parameter used to represent the actual number of subjects in the study.

A careful working definition is:

ACTSUB is a trial-level parameter representing the study’s documented actual subject count, as defined in the protocol, SAP, and metadata.

SDTM does not enforce a single rule for what must be included in ACTSUB. The exact population depends on how the study defines its “actual subjects.”

ACTSUB Is Not the Same as Treated Subjects

A common mistake is to assume ACTSUB equals treated subjects.

Population	Definition	Level
ACTSUB	Study-defined actual subject count	Trial
Safety Population	Subjects who received at least one dose	Analysis
ITT Population	Typically randomized subjects, per SAP	Analysis

These populations may overlap, but they are not required to be the same.

How ACTSUB Is Defined in Practice

In practice, ACTSUB reflects the study’s defined summary population. Depending on study design, this may align with enrolled subjects, randomized subjects, or another sponsor-defined population.

Because of this, ACTSUB should always be interpreted based on the study’s documented definition rather than assumed to follow a fixed rule.

Are Screen Failures Included in ACTSUB?

Generally no, but not by rule.

In many studies, screen failures are excluded from ACTSUB because the parameter reflects the enrolled or randomized population. However, this is not mandated by SDTM.

Screen failures are generally not included in ACTSUB when the parameter reflects enrolled or randomized subjects, unless the study’s documented definition specifies otherwise.

How Are Screen Failures Represented in SDTM?

Screen failures are primarily represented in the DM domain. Disposition details may also be captured in DS if collected and submitted.

Under SDTMIG guidance:

If a subject is not assigned to an arm, ARMCD is typically null and ARMNRS is populated.
ARMNRS is used to describe cases such as SCREEN FAILURE, NOT ASSIGNED, ASSIGNED, NOT TREATED, or UNPLANNED TREATMENT.

For scenarios where treatment does not align with planned arms, SDTM provides variables to describe actual treatment received, based on study mapping and SDTMIG guidance.

ACTSUB vs ARMCD

ACTSUB is a trial-level count, while ARMCD is a subject-level assignment variable.

They should not be treated as equivalent.

A subject may be included in ACTSUB even if ARMCD is missing, depending on study definition.
ARMCD alone should not be used to derive ACTSUB without understanding the study’s counting rule.

Edge Cases

Scenario	Included in ACTSUB?	Common DM Handling
Screen failure	Usually no	ARMCD null, ARMNRS = SCREEN FAILURE
Enrolled but not assigned to arm	Depends on study definition	ARMCD null, ARMNRS = NOT ASSIGNED
Randomized but not treated	Depends on study definition	Planned arm typically populated; ARMNRS clarifies no treatment
Unplanned treatment	Yes	ARMNRS and related variables describe situation

Do You Need a Screened Count in TS?

There is no universally required TS parameter for total screened subjects.

If a screened count is included in TS, it is typically a sponsor-defined parameter and should be clearly documented in metadata.

Note on Older Implementations

Older studies may use different approaches to represent screen failures and unassigned subjects. Current SDTMIG guidance emphasizes the use of ARMNRS and related variables, but implementation details may vary across studies.

Summary

ACTSUB is a TS parameter representing the study’s documented actual subject count.
It should not be assumed to mean treated subjects.
The exact population included depends on the study protocol, SAP, and metadata.
Screen failures are usually excluded, but not by strict SDTM rule.
DM variables (such as ARMCD and ARMNRS) handle subject-level assignment details.

References

CDISC SDTMIG guidance
CDISC Controlled Terminology
FDA Study Data Technical Conformance Guide

STUDYSAS BLOG