We have a situation where we want to detect and ignore documents/records at the earliest possible point based on certain criteria.  We have an existing "CDA Preprocessor" in the workflow which is an XSL transform.  This is what our contractors are planning to utilize to "filter" these documents.  This doesn't really make sense to me.  I see XSLT as a way to manipulate the _contents _of a document and filter certain peices of data.  In my brain, I would use an Ensemble business rule to detect the criteria and route the document as necessary. Does anyone have any input on this?  Our contractors have much more experience than I do with Healthshare, so I'd like to know why this makes more sense.