A Short History of CDISC and SAS Transport Files

Background

  • When development of the SDTM and SDTMIG started, SAS was in almost universal use in the pharmaceutical industry and at FDA.
  • "The SAS® Version 5 (V5) transport file format is an open standard developed by SAS to support data transfers between systems, especially those running different operating systems." SAS V5, being an open standard, allowed FDA to specify it as the standard required for data submission.
  • At that time, the CDISC volunteer team who developed the SDTMIG was focusing on drug submissions to FDA, so it was assumed that SDTMIG had to work with the SAS V5 transport file format. This assumption led to the creation of rules, such as a maximum variable name length of 8 characters, a maximum variable label length of 40 characters, and a maximum variable value length of 200 characters.

Current Regulation

  • FDA and PMDA require SAS V5 transport file format, and other regulatory agencies accept it.
  • If other recipients agreed, SDTM-based data could be exchanged in another file format. If both parties agreed, some of the restrictions rooted in the SAS V5 transport file format could be relaxed.

Changing Technology

  • Over the years, the CDISC user community has expanded, and although SAS is widely used for statistical programming in that wider community, other languages, such as R, are also used.
  • Non-tabular data formats, such as XML and RDF, are seeing greater use.

Responses to Changing Technology

  • CDISC has developed other formats, which could be used to represent SDTM-based data:
  • FDA has considered other formats and held a public meeting 5 Nov 2012 during which other alternatives (XML, RDF, newer SAS transport formats, etc.) were presented.
  • As a result, CDISC developed Dataset-XML and SDTM in RDF and conducted some testing in collaboration with FDA.   

Future Regulation

  • Regulatory agencies could require a different file format in the future.
  • Other parties who exchange SDTM-based data could decide to require a different exchange format.  If they did, and if demand were great enough, CDISC could develop implementation advice for that different exchange format.