Last year I had successful chats with people at IHTSDO about creating a very, very small subset of SNOMED that could be released under fair use (with that subset being approved by their legal department) so that it could be copied more freely without click-though agreements and the like. As I understand, what is important from IHTSDO’s perspective is that the actual codes would be incomplete and not usable in a production capacity. The purpose of the subset would be to demonstrate computational methods on SNOMED hierarchies so that regression tests, tutorials, and other materials could be produced.
At that time, we had discussed that such a tiny subset could be released with a statement in the NOTICE file for Apache with copyright and disclaimer about use, etc. The Apache license requires that the NOTICE file is maintained and not removed; typically it’s used for other copyright statements and such.
It seems that the example codes from the BookOfOHDSI and any small test data set that supports this book would make a fine example of such a limited data set. If we could discuss a set of codes and put together a proposal to IHTSDO, we could move on with getting this regression-test subset approved.