An Introduction to Therapeutic Area Standards

WORK IN PROGRESS

T. A. U. G.

CDISC data standards tend to have long titles, so each standard also has a short name, which is also used as its file name when published in PDF. Therapeutic area standards' short names always begin with the letters "TAUG". Why is this?

It has to do with CDISC's naming strategy. The final letter, "G", stands for "guide", because the document's purpose is to function as a how-to. The second-to-last letter, "U", stands for "user", which indicates the target audience. The first two letters, "TA", stand for "therapeutic area", because the document focuses not on a specific foundational standard, but on a specific therapeutic area.

In short: “TAUG” stands for “therapeutic area user guide”, an informative data standard that focuses on a specific therapeutic area.

What to Expect from a TA Standard

A TA standard typically provides advice and examples for the Clinical Data Acquisition Standards Harmonization model (CDASH), the Study Data Tabulation Model (SDTM), and/or the Analysis Data Model (ADaM), such as:

• Annotated sample case report forms (CRFs) compliant with CDASH

• CDASH metadata associated with the sample CRFs

• Guidance on the use of CDASH and SDTM variables

• Guidance on which domain models and datasets from the SDTM Implementation Guide for Human Clinical Trials (SDTMIG) to use in storing raw/collected data

• Examples of SDTM datasets, with text describing the situational context and pointing out records of note

• Cross-implementation variable definition metadata for non-standard (supplemental) variables used in example SDTM datasets and/or CRF mapping annotations

• Analysis datasets compliant with ADaM, with dataset- and variable-level metadata

• Table shells, mock reports, and diagrams illustrating the kinds of statistical analysis that can be performed based on the ADaM datasets

• Biomedical concept metadata exported from the Shared Health And Research Electronic (SHARE) metadata repository

A TA standard may also include:

describe common kinds of data needed for area studies, so that those handling the data (e.g., data managers, statisticians, programmers) understand the data and can apply standards appropriately. These descriptions sometimes include the clinical situations from which the data arise, and the reasons the data are relevant to the TA.

Does not include

•  Advice on what data to collect or how to analyze it
•  Information and advice already included in the foundational standards
•  Definitive controlled terminology (see the NCI-EVS website for the latest CT)
•  Implementation advice and terminology for questionnaires, ratings, and scales
•  Regulatory guidance or advice (though it includes some relevant references)
•  Clinical advice (though it includes some relevant references)

Does include

•  Advice on how to represent data using CDISC standards (but not those well covered in foundational standards – see above)
•  Examples illustrating this advice
•  Explanations for why the standards were implemented as shown in the examples, including clinical background relevant to modeling decisions
•  As part of those explanations, diagrams (concept maps) that illustrate clinical processes and/or relationships among data items
•  Where applicable, links to proposed additions to the standards (e.g., proposed new domains or variables) used in the examples
•  Links to Biomedical Concepts for some core TA data (these are maintained in SHARE, which is the authoritative source)

???

User guides do not replace the foundational CDISC standards or their implementation guides. Users should read those standards and implementation guides before applying the advice in user guides.

CDISC data standards are living documents. Due differing update cycles, some of the modeling approaches and controlled terminology presented in the examples in a document may become outdated before the next version of the document is released.

When a particular type of data has existing CDISC standards that can be used without additional development or customization, it is not covered in special detail in subsequent standards.

References

Clinical guidelines, articles, and other works consulted by the team during the creation of a document are referenced where appropriate. Documents that cite or rely on a large number of references include the full list of references in an appendix labeled "References".

Controlled Terminology

Concept Maps

CDISC often uses concept maps to explain clinical processes and biomedical concepts. Concept maps, also sometimes called mind maps, are diagrams which include “bubbles” representing concepts/ideas/things and labeled arrows that represent the relationships between the concepts/ideas/things. They are generally easier to draw and more accessible than more formal modeling diagrams, such as Unified Modeling Language (UML) diagrams.

The diagrams in CDISC standards use the following coding for classification of concepts:

This classification is based on classes in the Biomedical Research Integrated Domain Group (BRIDG) model (available at http://bridgmodel.nci.nih.gov/). These color-symbol pairs have been used to highlight kinds of things that occur commonly in clinical data and therefore give rise to common patterns of data. Some concepts are not coded; they have a thinner, black outline, and no accompanying symbol. These may include the subject of an observation, as well as characteristics, or attributes, of the coded concepts.

What is CFAST?

What is a "biomedical concept"?

A biomedical concept (BC) is a unit of knowledge, created by a unique combination of the characteristics that define observations of real world phenomena in clinical research and/or healthcare, which represents biomedical knowledge that borrows from medical knowledge, statistical knowledge, BRIDG, and the CDISC standards. Metadata for biomedical concepts include the properties of the data items that are parts of the concepts, controlled terminology for those data items, and the ways in which the concepts relate to each other.