In one of my previous entries, I showed how some SDTM variables are essentially superfluous as they are derived and the derivation can as well be done by the tool that reads the data, such as a viewer. Such features were already implemented in the "Smart Dataset-XML Viewer" (freely available from SourceForge). Using this viewer, the user can get values for all "--DY" variables calculated "on the fly", as well as the values for RFXSTDTC (first date of study treatment exposure) and RFXENDTC (last date of study treatment exposure) which are automatically retrieved from the "EX" dataset and displayed in the "DM" dataset.
One of my students (Wolfgang Hof) has taken the idea a step further and has developed a web service (using RESTful web service technology) for retrieving information about test codes (like LBTESTCD, VSTESTCD, ...). He developed a simple database and server software, and also generated a few test clients. He also implemented the service in the "Smart Dataset-XML Viewer". So if the user hovers the mouse over a cell containing a LBTESTCD value, the web service is triggered, a request send to the server, and the received information about the test code is displayed as a tooltip:
Now, what are the consequences of this?
First of all, it means that we could get rid of the variable "LBTEST", as there is a 1:1 relationship between LBTESTCD and LBTEST and the web service takes care of retrieving the value of "LBTEST" anyway. So essentially we could remove the variable LBTEST from the "required" variables in the SDTM-IG, also meaning that our LB submission files will decrease in size by something like 10-20%. That is surely something the FDA will like.
But take it another step further and imagine that we could query SHARE using a web service asking for additional information about the "CHOL" test code, returning information like which are the "preferred standardize units", what the FDA regards as being the normal range, and much much more, and use that information in the viewer, like marking values the FDA regards as being "out of range". Also SHARE already "knows" that "CHOL" is a test in the category "chemistry" so that we also can get rid of LBCAT.
A review environment using web services like the one our student Wolfgang Hof developed and using SHARE, wouldn't this be a great step forward?