Page History

These SDTM datasets represent the data collected above. An abbreviated Tobacco Product Identifiers and Descriptors (TO) domain example is also provided for reference.

These product malfunctions and events are represented in EM. Events not specifically associated with the product device should not be represented in this domain.     

The applicant used a standardized or dictionary-derived text for the description of an issue or events. (Note: CDISC does not prescribe what standardized or dictionary-derived text should be used.) The applicant decided to use the International Medical Device Regulators Forum (IMDRF) document terminologies for categorized adverse event reporting for the medical problem codes. The name and version of the dictionary used to map terms must be provided in a Define.XML ExternalCodeList element (see Section 2.10, Standards for Data Exchange).

EMDECOD was the Level 3 IMDRF term. NSVs were used to represent the various coding level variables defined in the dictionary. 

Example SUPPEM Variable Metadata

Example External Dictionaries

The SDTM EM dataset is shown below. 

The SUPPEM dataset is used to represent the NSVs used for coding the malfunction or event. The parent domain (RDOMAIN) is EM, and IDVAR is EMSEQ. QNAM holds the name of the supplemental qualifier variable being defined. This only shows the data for subject 2029. The data recorded in QVAL applies to the subject’s records, where IDVAR (EMSEQ) equals the value specified in IDVARVAL. 

In this study, subject 3067 had an AE associated with the battery problem. This was reported in the AE dataset. Not all variables are shown below for brevity.  

The RELREC dataset was used to record the relationship between the EM dataset and any reported AE represented in the AE dataset. 

Device events in tobacco studies may consist of device problems (which may or may not result in adverse events), device-reported warnings or alarms, calibration events, and replacement of parts. This information is represented in the Device Events (DE) and associated domains. In device studies, cases where the device did not perform as expected are typically called "events" or "incidents," rather than "problems" or "malfunctions," because often the true cause of the issue cannot be determined until a cause analysis is performed. This may or may not be a concern for a given trial depending upon whether the device has already been approved by regulators. Typically, post-approval device studies are less concerned about root-issue attribution.

There is more than one approach to identifying devices in studies. The method chosen will depend upon the granularity at which the sponsor needs to track the devices and will affect how the data are modeled. A device can be identified as a single unit (e.g., a syringe), or its components can be separately identified (e.g., barrel, plunger, needle). The level of granularity a sponsor chooses will be influenced by whether the components will be replaced and/or tracked, and how device/adverse event relationships and actions taken will be assessed.