CDISC ADaM standard

ADaM is a CDISC standard to submit analysis data to FDA. Key is to understand that ADaM is a SDTM model for providing analysis data, programs and data definitions. The principles of ADaM is aimed at providing  a clear and unambiguous  communicationof the content, source and  quality of the datasets submitted in support of  the statistical analysis performed by the 
sponsor. This in turn would support the machine-readable  description for the JANUS data repository. 

Analysis datasets

Analysis datasets are datasets created to support specific analyses.

• Each dataset is provided as a SAS Transport (XPORT) file. 
• Programs should be provided as both ASCII text and PDF files and should include sufficient documentation to allow a reviewer to understand the submitted programs. 

Programs

Programs are scripts used with selected software to produce reported analyses based on these datasets.

Analysis-level Metadata

• ANALYSIS NAME –A unique identifier for this analysis. May include a table number or other sponsor- specific reference. 
• DOCUMENTATION –A text description documenting the analysis performed.
• REASON –The reason for performing this analysis. Examples may include Pre-specified, Data-driven, Exploratory, and Regulatory Request. 
• DATASET –the name of the analysis dataset used should be linked to the analysis dataset used for this analysis. In most cases, this will be a single dataset. If multiple datasets are used, they should all be listed here. 
• PROGRAM –Analysis programs using the DATASET above as input can be described or included here.

The complete ADaM specification can be found at : http://www.cdisc.org/models/adam/V2.0/index.html

define.xml and SDTM

Define.xml is the document which specifies the standard for providing Case Report Tabulations Data Definitions in an XML format for submission to regulatory authorities (e.g., FDA). 

The XML schema used to define the expected structure for these XML files is based on an extension to the CDISC Operational Data Model (ODM).

In the picture below the key aspects are highlighted with the red text, basically it is a listing of data items that a sponsor will transmit to FDA along with the regulatory application.

Each data set if furthered detailed with what data items are sent and their nature (ex: Derived or from a CRF page)

You can find the DTD definition and stylesheets to transform define.xml on CDISC website : http://www.cdisc.org/models/def/v1.0/index.html

CDISC ODM Standard

The Operational Data Model (ODM) provides a format for representing the study metadata, study data and administrative data associated with a clinical trial. 

It represents only the data that would be transferred among different software systems during a trial, or archived after a trial. 

It need not represent any information internal to a single system, for example, information about how the data would be stored in a particular database.

The ODM model assumes that a study’s clinical data will consist of several kinds of entities. These include: 

• subjects
• study events (a series of forms connected to an event)
• forms (aggregations of item groups)
• item groups (groups of items that will be analyzed together)
• items (single data item such as Hb%)
• annotations (comment applied to any of the above)

CDISC SDTM standard

Inline with the Critical Path Initiative CDISC has developed three main standards to improve the efficiency of clinical trial projects, namely : Operational Data Model (ODM), the Study Data Tabulation Model (SDTM) and the Analysis Data Model (ADaM).

The SDTM defines a standard structure for study data tabulations. These  are to be submitted as part of a product application to a regulatory authority such as FDA. 

The SDTM was prepared by the CDISC Submission Data Standards (SDS) Team to guide the organization, structure, and format of tabulation data sets for study data submitted to regulatory authorities. Data tabulation data sets are one of four ways to represent the human subject Case Report Tabulation (CRT) and equivalent animal data submitted to the FDA.

The SDTM is composed of 30+ defined domains within six broad categories. Each domain represents a file structure and contains a particular type of data associated with clinical trials, such as demographics, vital signs or adverse events. 

The model also provides the ability to create custom-defined domains with sets of standard variable definitions. Variables in common across domains all have similar name extensions, and the standard specifies the beginning prefix of all variables be a (typically) two-letter domain abbreviation.

We will come back with a discussion on how to interface your CDM system to produce SDTM files.

CDISC discussion forums : http://www.cdisc.org/discussions/discussions.html

CDISC

CDISC mission statement from CDISC website says “CDISC is a global, open, multidisciplinary, non-profit organization that has established standards to support the acquisition, exchange, submission and archive of clinical research data and metadata. The CDISC mission is to develop and support global, platform-independent data standards that enable information system interoperability to improve medical research and related areas of healthcare.  CDISC standards are vendor-neutral, platform-independent and freely available via the CDISC website.“

We will cover each CDISC standard in detail in the coming posts, here is a summary table with links to the current standard specification at http://www.cdisc.org. To give an overview the different parts of the CDISC standard are related as below to the clinical trial process

CDISC Standards in Production

Submission Data Standards Team (For submission of data-sets to regulatory)	(SDTM IG V3.1.1) (SDTM V1.1) (SDTM IG V3.1) WebSDM edit checks for  (SDTM 3.1.1)
Operational Data Modeling Team	(ODM V1.3) (ODM V1.2.1) See also (eSDI Document)
Analysis Dataset Model Team	(ADaM 2.0)
Laboratory Standards Team	(LAB)
Standard for Exchange of Non-clinical Data	(SEND V2.3)
Case Report Tabulation Data Definition Specification (define .xml)	(CRT-DDS V1.0)
Terminology	(Terminology)

Standards in Development

Submission Data Standards Team	SDTM IG V3.1.2 Draft SDTM V1.2 Draft Metadata Submission Guidelines, Appendix to the SDTM IG V3.1.1
Protocol Representation Group	(PRG)
Clinical Data Acquisition Standards Harmonization (CDASH)	(CDASH)
Terminology	(Terminology)
Cardiovascular and Tuberculosis Data Standards	(Cardiovascular and Tuberculosis Data Standards)

Data integration

Data Integration has become a major (and unsolved) challenge in the realm of clinical trials. As the means to electronically capture and store the data from clinical trials in large databases became easy the data started pouring in from several sources and different formats.

For example, the use of third party applications like IVRS, ePRO, CDMS, CTMS and AERS systems is a great challenge since they could come from different vendors and use their own internal proprietary data formats. 

Another key area is laboratory data integration. Often the lab data comes from a LIMS or a data dump from the lab. This data has to be integrated with rest of the trial data. There is also a temporal dissociation between clinical data and lab data.

Often the issue becomes more challenging due to anonymization or coding. Each patient has to be identified by the same unique identifier across different systems to make integration feasible.

In the next couple of posts we will discuss about CDISC and the emergence of standards to support data integration.