Study Subject Data

Observations about study subjects are normally represented in a series of domains. A domain is defined as a collection of logically related observations with a common topic. The logic of the relationship may pertain to the scientific subject matter of the data or to its role in the trial. For example, "Subject 101 had an adverse event of mild nausea starting on study day 6" is an observation belonging to the Adverse Events (AE) domain in a clinical trial.

The entity that is the subject of a study varies across studies. The SDTM variable that represents an observation about a study subject or a group of study subjects can be

• USUBJID for observations about individual humans or animals who are study subjects.
• POOLID for observations about a group of human or animal study subjects which cannot be attributed to individual subjects.
• SPDEVID in certain domains for a study in which devices are subjects of the study.
• SPTOBID for observations about tobacco products which are subjects of the research.
  
• GTREFID for observations about sources of information that serve as subjects in genetic toxicology testing performed in vitro (GT domain only).

SDTM datasets that represent data about study subjects are of 2 types: general observation class datasets and special-purpose datasets. General observation class domains conform to general structures for 1 of 3 classes: findings, events, and interventions. The data for a study would generally include multiple domains in each general observation class. In contrast to general observation class models on which individual domains can be based, special-purpose domains do not to conform to a common structure and may make use of variables that do not appear elsewhere in SDTM.

Domains based on the general observation classes are specified in SDTM implementation guides (see Section 1.2, Implementation Advice for this Model).

Each study subject domain dataset is distinguished by a unique 2-character code. This code, which is stored in the SDTM variable named DOMAIN, is used in the dataset name, as the value of the DOMAIN variable in that dataset, and as a prefix for most variable names in that dataset.

Domain codes are also used in RDOMAIN, a variable that is included in several relationship datasets and in the Comments special-purpose domain.