SDTM


An Introduction to CDISC Data Standards

Tags: | | | |

data1CDISC is the Clinical Data Interchange Standards Consortium, a nonprofit organization whose goal 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 develops standards for packaging and reporting clinical research and related data. According to the CDISC website, use of these data standards will “become mandatory 2 years after the final guidance is issued for NDAs and BLAs, and will apply to all new studies begun 1 year after the final guidance is published.” The final guidance is required by fiscal year 2017.

Some of the more significant standards include:

CDASH (Clinical Data Acquisition Standards Harmonization) is used for acquiring clinical study data. The goal of CDASH is to harmonize clinical element names, definitions and metadata.

SDTM (Study Data Tabulation Model) is used for reporting clinical study data in a machine-readable format. SDTM reports clinical observations as series of variables. These variables represent values in a corresponding table.

ADaM (Analysis Data Model) is used for analyzing data.

LAB (Laboratory Data Model) is a standard for transmitting laboratory data. Unlike the other standards, the name is not an acronym. It was originally created to facilitate transferring data from Labs to Clinical organizations.

The different standards are used for different parts of the clinical process: acquiring data (CDASH), reporting data (SDTM) and analyzing data (ADaM).

The usual flow of clinical data through the CDISC standards is:

  1. Data are collated from medical source documents and collected in CDASH formats.
    If there is clinical laboratory data, this can be concurrently collected in LAB format and added to CDASH datasets.
  2. CDASH and LAB data are transcribed to SDTM format for reporting.
  3. SDTM data are reformatted and updated to ADaM datasets.
    (Some organizations go directly from CDASH to ADaM datasets.)
  4. ADaM datasets are analyzed.
  5. Study tables and listings are generated using data lists from the CDASH or SDTM datasets and the analysis results are generated from the ADaM datasets.

Data sets are generally only human-readable at the beginning and end of this process. The CDASH and LAB can be human readable when viewed directly with some experience.  SDTM and ADaM data are not human-readable; programs use SDTM and ADaM data to produce the tables, listings, figures, and statistical results that are the core of the Clinical Study Report (CSR).