It's a packet with multiple documents in one PDF.  The first page has a QR code with the employee number and some other metadata.  The first doc class in the QF session reads the bar code on the first page and stores the metadata in token collectors.  The following pages are then identified with OCR as different doc classes, but retain the metadata using token retrievers.  Since we scan one packet at a time, I want to be certain that there is metadata in the token collectors before running the subsequent doc classes.

You might be able to do something using two separate Quick Fields sessions.  The first session would simply check for the barcode information and, if present, set that information as fields and automatically store that "packet" in a temporary folder.  If the barcode information was missing, the document would be unidentified and stay in the session until a user filled it in manually and stored them.

The second session would pick up documents from the temporary folder and break them up, using the fields containing the barcode information.