4 Radnor Corporate Center
Suite 350
Radnor, PA 19087
+1 (888) 507-2270
SD1057, SD1058, SD1059
Hi Randall, There are a lot
Hi Randall,
There are a lot of discussions around those new rules. Please see http://www.opencdisc.org/forum/non-recommended-variable-length
Do you have any comments on that?
Thank you,
Sergiy Sirichenko
OpenCDISC versus FDA versus SDTM
Sergiy,
If OpenCDISC is going to check lengths, then the lengths should be maximums, not set values. This is the only way to comply with FDA's variable-length directive (right?). The FDA directive introduces several potential issues, such as the risk of truncation when different study data are combined, but those will have to be managed. It doesn't seem that OpenCDISC should guide people away from satisfying FDA's length request, especially in the absence of an SDTM rule about length (other than maximums).
Variable lengths - inconsistent FDA guidance
What a crap!
The FDA tells us we should optimize our datasets so that they do not become larger than absolutely necessary, and when we do so, OpenCDISC gives us a warning that we MUST set the variable length to the maximum that is allowed.
About the risk of truncation when combining different datasets, that is a software issue, not a data issue. Good software can handle such situations very well, crap software can't.
Are we really still living in the era of punchcards?
Rules SD1057-SD1059 - Non-recommended variable length for certain TEST and related variable values - expects the variable lengths to equal the max allowable. This is inconsistent with the "CDER Common Data Standards Issues" request for field lengths to be adjusted to the max found in the field, in order to reduce file sizes. Rather than testing whether the field length equals the max, shouldn't the check be that the field length doesn't exceed the max?