XML files

With the advent of the CDISC ODM model and the progression of the FDA s endorsement of the CDISC models, I believe that eventually all clinical trial data will likely be submitted to the FDA in ODM or a similar XML format. The XML-based ODM is already gaining acceptance within the pharmaceutical industry as a means of transferring clinical trial data. SAS provides two ways to produce ODM data files using either PROC CDISC or the XML LIBNAME engine. 

PROC CDISC is a SAS procedure available as a hot fix for SAS 8.2 that ships as part of SAS 9.1.3. PROC CDISC allows you to export (and import) XML files that are compliant with the CDISC ODM version 1.2 schema. Here are a sample PROC CONTENTS and a sample PROC PRINT of a SDTM data set called DM (for demographics) that we will export to ODM using PROC CDISC. 

16 rfendtc Char 16 Subject Reference End Date/Time 

6 rfstdtc Char 16 Subject Reference Start Date/Time 

0 bs studyid usubjid siteid subjid do main rfstdtc brthdtc dmdtc age ageu dm dy country arm arm cd se X rfendtc race 

---

In most cases, the data that you use for clinical trial analyses are found in some kind of computer file external to the SAS System. The data you need may be found in a permanent SAS data set, a relational database table found in Oracle or Microsoft SQL Server, a Microsoft Access or Excel file, a simple delimited ASCII text file, or even an XML file. In any case SAS provides a wide array of ways in which to import data files into SAS. We explore these tools and the advantages and disadvantages of each in this chapter. 

On occasion you may find that you need to import data from ASCII text files for analysis. In this section, traditional rectangular ASCII text files are discussed. Although XML files are composed of ASCII text, they pose unique challenges that are discussed later in this chapter. Some examples of ASCII text data that you might need to import include... 

Because the data are stored as simple ASCII text, text file browsers far into the future should still be able to read an XML file created today. 

Beginning with SAS 8, there are tools in SAS to help you read and write XML files. Starting with SAS 9.1, SAS provides XML import support via the XML LIBNAME engine (previously called SXLE) and the SAS XML mapper (previously called Atlas). 

There are two things that you may notice when looking at the lab normal data represented as XML. First, the file seems verbose. Whereas previously the lab normal file could be represented with three lines of pipe-delimited text, XML represents the same data with 30 lines of text. Second, you can read and somewhat understand the XML file just by looking at it if you know a markup language such as HTML or SGML. Let s look at how we can import these XML data into SAS. 

Because the XML map file is valid XML itself, you can read how the SAS variables are translated from the XML lab normals data file. Once the XML map file is defined, you just need the simple SAS program that follows to read the lab normals XML file into SAS. 

XMLMap Settings tab Allows for XML file format encoding to be set and allows for... 

XML will become more integral to the work of statistical programmers in the pharmaceutical industry as the standards, applications providers, and vendors make more use of this technology. Eventually you should expect the FDA to move away from SAS transport files to XML files as their standard data format for electronic data submission. 

Program 8.3 Using PROC CDISC to Create an ODM XML File... 

The subsequent slightly truncated dm.xml file looks like this ... 

Complete SAS documentation for PROC CDISC can be found at http //support.sas.com/md/base/topics/sxle82/TW8774.pdf. In the future, it is expected that PROC CDISC will be able to export multiple SAS data sets to a single ODM XML file. As the CDISC models and PROC CDISC rapidly evolve, continue to watch PROC CDISC as a tool for CDISC model conversion. 

The resulting ODM XML file produced by the XML LIBNAME engine is very similar to the one produced by PROC CDISC. However, PROC CDISC allows you to specify and pass more of the metadata information along to the XML file. If you need to create customized XML files for a sponsor that do not match the ODM specification, you can use the SAS XML Mapper and the XML LIBNAME engine to write your own custom XML files. 

---