Being neck deep in clinical data ETL right now, I would love to hear some details on how one does real-time ETL and reporting while leaving data in place. Most of the clinical systems I've worked with have some pretty ugly data structures under the hood that are just barely relational.
For a simple example: they might take a question that is a yes/no question and store the answers on two rows in the database: one for yes (with a value of 1 or 0) and one for no (also with a value of 1 or 0).
I long ago gave up on the thought that one could leave the data in place and work on it (vs traditional data warehousing approaches). If these guys have solved that problem, that's a product all on its own.
Thanks for your comment. I'd be happy to talk about it with you more, and potentially setup a demo against your data. If you're interested, please email me at my username at comprehendclinical.com, or fill out this form: http://www.comprehendclinical.com/demo.html
I long ago gave up on the thought that one could leave the data in place and work on it (vs traditional data warehousing approaches). If these guys have solved that problem, that's a product all on its own.